I think your latter point is closer to the truth. A similar example: in the UK about 20% of girls and 40% of boys study physics at A-level (the final school ex

I think your latter point is closer to the truth. A similar example: in the UK about 20% of girls and 40% of boys study physics at A-level (the final school exam before university). Every year everybody seems surprised that the girls outperform the boys. However, it is likely that the students who study physics are mostly the students with the greatest aptitude for the subject. If you split the students into any two arbitrary groups then the top fifth from one group are going to outperform the top two fifths from the other. This is unremarkable. Yet when the groups are boys and girls, people are surprised. Anyway, I would be surprised if something similar isn't happening in CS. With so few women studying the subject, it's likely that those who do are fairly dedicated.